LAMP3) Enhances

European Journal of Cell Biology 94 (2015) 148–161

Contents lists available at ScienceDirect

European Journal of Cell Biology

jo urnal homepage: www.elsevier.com/locate/ejcb

Loss of lysosome-associated membrane protein 3 (LAMP3) enhances

cellular vulnerability against proteasomal inhibition

a a a

Jorge Antolio Dominguez-Bautista , Michael Klinkenberg , Nadine Brehm ,

b b b a

Mahalakshmi Subramaniam , Beatrice Kern , Jochen Roeper , Georg Auburger , a,∗

Marina Jendrach

Experimental Neurology, Department of Neurology, Goethe University Medical School, Frankfurt am Main, Germany

Institute of Neurophysiology, Goethe University Medical School, Frankfurt am Main, Germany

a r t i c l e i n f o a b s t r a c t

Article history: The family of lysosome-associated membrane proteins (LAMP) includes the ubiquitously expressed

Received 25 August 2014

LAMP1 and LAMP2, which account for half of the proteins in the lysosomal membrane. Another member

Received in revised form

of the LAMP family is LAMP3, which is expressed only in certain cell types and differentiation stages.

26 December 2014

LAMP3 expression is linked with poor prognosis of certain cancers, and the locus where it is encoded was

Accepted 15 January 2015

identiﬁed as a risk factor for Parkinson’s disease (PD). Here, we investigated the role of LAMP3 in the two

main cellular degradation pathways, the proteasome and autophagy. LAMP3 mRNA was not detected in

Keywords:

mouse models of PD or in the brain of human patients. However, it was strongly induced upon proteaso-

Autophagy

Cancer mal inhibition in the neuroblastoma cell line SH-SY5Y. Induction of LAMP3 mRNA following proteasomal

LAMP3 inhibition was dependent on UPR transcription factor ATF4 signaling and induced autophagic ﬂux. Pre-

Lysosome vention of LAMP3 induction enhanced apoptotic cell death. In summary, these data demonstrate that

Parkinson’s disease LAMP3 regulation as part of the UPR contributes to protein degradation and cell survival during protea-

Proteasome somal dysfunction. This link between autophagy and the proteasome may be of special importance for

the treatment of tumor cells with proteasomal inhibitors.

BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Introduction and Ciechanover, 2002). Autophagic-lysosomal degradation can be

classiﬁed into macroautophagy, chaperone mediated autophagy

The ubiquitin-proteasome system and the autophagic- and microautophagy. In macroautophagy (usually referred to as

lysosomal pathway are the two major degradation systems autophagy) the substrate cargo is engulfed by the autophagosome,

for proteins in eukaryotic cells. They are responsible for the degra- which is characterized by the presence of the protein LC3-II.

dation of unnecessary, dysfunctional or damaged components, Once the cargo is completely surrounded by the autophago-

from soluble proteins to protein aggregates and whole organelles some, this double-membrane organelle fuses with the lysosome,

(Korolchuk et al., 2010). The ubiquitin-proteasome system medi- where proteolysis occurs (He and Klionsky, 2009; Mizushima,

ates the enzymatic ubiquitination of substrate proteins. The tagged 2007).

proteins are then recognized by a barrel-shaped structure called The ubiquitin-proteasome system and autophagy do not act

the proteasome, in which they are degraded to peptides (Glickman independently from each other. Defective autophagy results in

accumulation of ubiquitinated proteins, impacting the ﬂux of the

ubiquitin proteasome system. On the other hand, dysfunction of the

proteasome can promote a compensatory induction of autophagy

Abbreviations: CCCP, carbonyl cyanide m-chlorophenyl hydrazone; DMSO,

(Korolchuk et al., 2010; Nedelsky et al., 2008).

dimethyl sulfoxide; FCS, fetal calf serum; GFP, green ﬂuorescent protein; HBSS,

Hank’s Balanced Salt Solution; LAMP, lysosome-associated membrane protein; LC3, Lysosomal function depends on lysosomal hydrolases and inte-

microtubule-associated protein light chain 3; PARP, poly ADP ribose polymerase;

gral lysosomal membrane proteins. Over 25 lysosomal membrane

PD, Parkinson’s disease; ROI, region of interest; UPR, unfolded protein response.

∗ proteins are known, which participate in diverse tasks such as lyso-

Corresponding author. Current address: Institut für Neuropathologie Charité -

somal acidiﬁcation, membrane fusion and transport of degradation

Universitätsmedizin Berlin Campus Charité Mitte Charitéplatz 1 | Virchowweg 15,

products to the cytoplasm (Saftig and Klumperman, 2009). The

D-10117 Berlin, Germany. Tel.: +49 30 450 536291.

E-mail address: [email protected] (M. Jendrach). family of lysosome-associated membrane proteins (LAMP) includes

http://dx.doi.org/10.1016/j.ejcb.2015.01.003

J.A. Dominguez-Bautista et al. / European Journal of Cell Biology 94 (2015) 148–161 149

ﬁve members: LAMP1, LAMP2, LAMP3, BAD-LAMP and macrosialin supplemented with 3% FCS and 10 ␮M retinoic acid for 10 days.

(CD68/LAMP4), which share a so-called LAMP domain and possess New medium was added every other day.

several N- and O-glycosylation sites (Saftig et al., 2010; Wilke et al.,

2012).

Infusion of mouse brains with epoxomicin

LAMP1 and LAMP2 represent about 50% of the lysosomal mem-

brane proteins (Saftig et al., 2010). They are type-1 transmembrane

Infusion of mice midbrains was performed as described else-

proteins with high sequence homology, containing a highly gly-

where (Subramaniam et al., 2014). Brieﬂy, 4 ␮l of epoxomicin

cosylated luminal domain and a short cytosolic tail (Saftig and

(10 ␮M) or DMSO (1%) were unilaterally injected into the substantia

Klumperman, 2009). The generation of LAMP1 knockout mice

nigra of 11-week old mice. Either 24 h or 2 weeks post-injection, the

resulted in mild regional brain astrogliosis and upregulation of

mice were euthanized. The midbrain (including substantia nigra) of

LAMP2 at the protein level. However, the distribution and den-

the injected side was immediately dissected, frozen in liquid nitro-

sity of lysosomes, as well as lysosomal enzyme activity, pH and ◦ ®

gen, and stored at −80 C. Total RNA was isolated using the TRI

osmotic stability were unaffected (Andrejewski et al., 1999). On

reagent, and qRT-PCR analysis was performed as described below.

the other hand, LAMP2 knockout mice showed increased mortality

between 20 and 40 days of age, with the surviving mice being fertile

Post-mortem samples of human brain

and having a normal life span, although they showed accumula-

tion of autophagic vacuoles in several tissues (Tanaka et al., 2000).

Brain tissue samples from autopsy-conﬁrmed subjects with

A LAMP1/LAMP2 double knockout resulted in embryonic death

Parkinson’s disease with Braak stage 6 (n = 6) of ␣-synuclein

between E14.5 and E16.5. Fibroblasts derived from those double

pathology and age-/gender-matched controls (n = 4, Supplemen-

knockout embryos showed accumulation of autophagic vacuoles

tary Table 1) were obtained from the Neurobiobank Munich with

and of LC3-II after amino acid starvation, indicating impairment of

approval of the Frankfurt University Medical school ethics com-

autophagic ﬂux (Eskelinen et al., 2004).

mittee. Pathological analysis was performed according to the Lewy

Unlike the ubiquitous LAMP1 and LAMP2, LAMP3 (also called

Body Dementia staging (Braak et al., 2003). At ages above 60 years,

DC-LAMP, CD208, or TSC403) is expressed only in speciﬁc tissues

all PD and control cases had some Alzheimer-associated neuroﬁb-

and conditions. Early reports on LAMP3 showed that it is induced

rillary tangle pathology (Braak and Braak, 1991). The PD cases had

upon human dendritic cell differentiation (de Saint-Vis et al., 1998),

conﬁrmed ␣-synuclein pathology in the cingulate gyrus, while the

that it is present in normal and transformed human type II pneu-

controls were chosen so that no ␣-synuclein-, amyloid-beta- or

mocytes (Akasaki et al., 2004; Salaun et al., 2004), and is increased

MAP-tau-pathology was detectable in this brain region (Thal et al.,

in carcinomas of different origin, where it has been linked to metas-

2002). Gray matter from the posterior cingulate gyrus next to the

tasis of tumor cells and poor prognosis (Kanao et al., 2005; Ozaki

posterior thalamus was dissected for total RNA and protein isola- et al., 1998).

tion.

Additionally, despite the lack of evidence for LAMP3 expression

Total protein was obtained from ∼150 mg of tissue by serial

in brain tissue (Akasaki et al., 2004; de Saint-Vis et al., 1998), several

isolation. First, the tissue was homogenized with a Dounce tissue

genome-wide association studies have identiﬁed the chromoso-

grinder (Wheaton) and 10 volumes of RIPA buffer (50 mM Tris–HCl

mal locus MCCC1/LAMP3 associated with increased risk for sporadic

pH 7, 150 mM NaCl, 0.1% SDS, 1% Triton X-100, 0.5% sodium

old-age Parkinson’s disease (PD) (Li et al., 2013; Lill et al., 2012;

deoxycholate, 2 mM EDTA, and 1× protease inhibitors). Lysate was

Pihlstrom et al., 2013). Finally, LAMP3 was shown to be involved in ◦

centrifuged at 9000 × g for 10 min at 4 C and the RIPA-insoluble

the unfolded protein response (UPR) during hypoxia (Nagelkerke

pellet was then lysed with 2× SDS buffer (137 mM Tris–HCl pH 6.8,

et al., 2013a), while its localization points to a possible function in

4% sodium dodecyl sulfate, 20% glycerol, 1× protease inhibitors),

autophagic-lysosomal degradation.

sonicated for 10 s with a homogenizer (Bandelin) and centrifu-

As recent data indicate that the autophagic-lysosomal pathway

gated at 9000 × g for 10 min. The 2× SDS-soluble fraction contains

and the ubiquitin-proteasome system act together to maintain pro-

membrane-associated proteins and was used for LAMP3 western

tein homeostasis (Korolchuk et al., 2010; Park and Cuervo, 2013; ®

blot. Total RNA was isolated from ∼50 mg of tissue using the TRI

Vogl et al., 2014), we evaluated the hypothesis that LAMP3 repre-

reagent. RNA integrity was determined using the Agilent RNA 6000

sents a link between these two systems.

Pico Kit and the Agilent 2100 Bioanalyzer (Agilent). Gene expres-

sion was measured by qRT-PCR as indicated below.

Materials and methods

Plasmid, siRNAs, and cell transfection

Cell culture

A plasmid coding for LAMP3 with enhanced green ﬂuorescent

protein fused at the C-terminal end (LAMP3-GFP) was gener-

HeLa cells were cultured in Minimal Essential Medium with

ated by subcloning the open reading frame of human LAMP3

Earle’s salts (Gibco) supplemented with 1× MEM Non-Essential

(NM 014398.3) into the XmnI and NotI restriction sites of vec-

Aminoacids (Gibco), 10% FCS Gold (PAA), and 10 mM HEPES buffer

tor pEZ-M03, under the control of the Cytomegalovirus promoter

10 mM (PAA). Neuroblastoma SH-SY5Y cells were cultured in

(GeneCopoeia).

RPMI 1640 medium (Gibco) supplemented with 10% FCS (PAA).

◦

The siRNAs AllStars Negative Control (1027281), Hs LAMP3 5

Both cell lines were kept at 37 C, 5% CO2 and 95% air. SH-SY5Y

(SI041532226), Hs LAMP3 7 (SI04287458), and Hs ATF4 5

cells were cultured with the following drugs at the indicated

(SI03019345) from QIAGEN were used.

ﬁnal concentration: U0126 10 ␮M, LY294002 25 ␮M, CCCP 10 nM,

The Amaxa Cell Line Nucleofector kit V and the Nucleofector

rapamycin 0.5 ␮M, staurosporine 30 nM, baﬁlomycin A1 10 nM, 3-

2b Device (Lonza) were used to transfect 1.5 × 10 SH-SY5Y

methyladenine 10 mM, MG132 3 ␮M or 300 nM, epoxomicin 1 ␮M,

cells with 0.1 nmol of siRNAs according to the manufacturer’s

thapsigargin 1 ␮M, and tunicamycin 1 ␮g/ml. DMSO was included

instructions. Transfection of HeLa cells was performed by plat-

as control at a ﬁnal concentration of 0.05%. Cells were deprived

ing 250 000 cells on gelatin-coated 24-mm coverslips. Cells were

of amino acids by culturing them in Hank’s Balanced Salt Solution

transfected using 1 ␮g of plasmid with the Effectene kit (QIAGEN)

HBSS (Gibco). Differentiation of SH-SY5Y cells was performed by

2 according to manufacturer’s instructions.

culturing cells at a density of 52 000 cells/cm in RPMI medium

150 J.A. Dominguez-Bautista et al. / European Journal of Cell Biology 94 (2015) 148–161

cDNA synthesis and quantitative RT-PCR the lysate was centrifuged at 9000 × g for 10 min. Supernatant was

collected and the protein concentration was determined by the

Total RNA from cells was isolated using the RNeasy Mini Bradford method using bovine serum albumin as reference.

Kit (QIAGEN) according to manufacturer’s instructions. One Between 1 and 5 ␮g of protein were mixed with caspase-3 reac-

microgram of total RNA was used for cDNA synthesis using the tion buffer (25 mM HEPES pH 7.4, 1 mM EDTA, 0.1% Chaps, 1 mM

SuperScript III Reverse transcriptase (Life Technologies). Quan- DTT, 10 ␮M AC-DEVD-AMC, and 8.6% d(+)-Sucrose). Fluorescence

titative Real Time Polymerase Chain Reaction (qRT-PCR) was intensity was measured every 10 min over the course of 2 h using

performed in a Step One Plus Real Time PCR System (Life Techno- excitation at 360 nm and emission at 465 nm in a GENios microplate

logies) using the following TaqMan assays: LAMP1 (Hs00931- reader (Tecan).

464 m1), LAMP2 (Hs00903587 m1, which detects the LAMP2A,

LAMP2B, and LAMP2 C isoforms), LAMP3 (Hs01111317 m1), ATG5

Confocal laser scanning microscopy and immunostaining

(Hs00355494 m1), BECN1 (Hs00186838 m1), ATG7 (Hs00197-

348 m1), DDIT3 (Hs00358796 g1), HSPA5 (Hs99999174 m1), ATF3

Cells were cultured on gelatin-coated coverslips, ﬁxed with

(Hs00231069 m1), ATF4 (Hs00909569 g1), SQSTM1 (Hs001776-

4% paraformaldehyde for 20 min, washed 3 times with phosphate

54 m1), PINK1 (Hs00260868 m1), PARKIN (Hs01038318 m1), ATP-

buffered Saline 1× (PBS 1×), permeabilized with 0.1% Triton X-100

13A2 (Hs01119446 m1), DJ-1 (Hs00697109 m1), SNCA (Hs011033-

for 20 min, washed 3 times with PBS 1×, blocked with 3% bovine

86 m1), ATNX2 (Hs00268077 m1), EIF4A2 (Hs01115195 g1), SYT11

serum albumin for 1 h, and incubated overnight with the respec-

(Hs01064643 m1), CPLX1 (Hs00362510 m1), NSF (Hs009380-

tive primary antibody: LAMP2/CD107b, which detects the three

40 m1), RAB7L1 (Hs01026316 m1), RAB25 (Hs01040784 m1), GAK

isoforms of LAMP2 (BD Biosciences), p62/SQSTM1 (Santa Cruz) and

(Hs01049227 m1), BST1 (Hs 01070189 m1), STK39 (Hs010853-

LAMP1 (BD Bioscience). Next day incubation with a Cy3-conjugated

46 m1), HIP1R (Hs00391321 m1), EIF4A2 (Hs01115195 g1),

secondary antibody (Jackson ImmunoResearch) was performed for

GAPDH (Hs99999905 m1), TBP (Hs99999910 m1), Lamp3

2 h and the nucleus was counterstained with Hoechst 33258.

(Mm00616604 m1), and Tbp (Mm00446973 m1). TBP or Tbp

Microscopic analysis was performed using a Leica TCS SP5 con-

were used as human or mouse housekeeping genes, correspond-

focal laser scanning microscope and a HCX PL APO lambda blue

−Ct

ingly. Relative gene expression was calculated using the 2

63× OIL UV objective controlled by LAS AF scan software (Leica

formula.

Microsystem, Wetzlar, Germany). Images were taken simultane-

ously and co-localization was determined with the program Fiji

Western blot

(function Coloc2) using single slices and ROIs. The number of lyso-

somes was determined with the program Fiji as described before

Total protein was isolated with 2× SDS lysis buffer (137 mM

(Parganlija et al., 2014). Images in the ﬁgures represent maximum

Tris–HCl pH 6.8, 4% sodium dodecyl sulfate, 20% glycerol, 1× pro-

image projections.

tease inhibitors). The cell lysate was sonicated for 10 s with a

homogenizer (Bandelin) and centrifugated at 9000 × g for 10 min.

Data analysis

Supernatant was collected and protein concentration was deter-

mined with a protein quantitation kit (Interchim) using bovine

Experiments were repeated at least 3 times independently. Val-

serum albumin as standard. Human lung tissue lysate was obtained

ues are presented as mean ± SEM (Standard Error of the Mean).

from Novus biologicals. 15–30 ␮g of total protein lysate were

Statistical difference between means was assessed by the two-

separated by SDS-PAGE, transferred to a PVDF membrane, and

tailed Student t-test when comparing two groups, and by one-way

blocked with 5% skim milk in 1× PBS/0.1% Tween. For analy-

ANOVA with Bonferroni’s post hoc test when comparing three

sis of brain tissue, 15 ␮g of protein were separated using 4–12%

or more groups using the GraphPad Prism software. Statisti-

Bis–Tris pre-cast gels (Novex). Membranes were probed using

cally signiﬁcant values are indicated by *p ≤ 0.05, **p ≤ 0.01, and

the following primary antibodies: LC3 (Sigma), SQSTM1 (Santa

***p ≤ 0.001.

Cruz), PARP (Cell Signaling), GFP (Living Colors), and GAPDH (Cal-

biochem). The following LAMP3 antibodies were used: ab83659

(Abcam), AF4087 (R&D systems), MABC44 (Millipore), and 10527- Results

RP02 (Sino Biological Inc.). Secondary antibodies were conjugated

with horseradish-peroxidase, and detection was performed with LAMP3 is detectable in the neuroblastoma cell line SH-SY5Y but

West pico chemiluminescent Super Signal Substrate. Band intensity not in brain

was quantiﬁed using the ImageJ software.

Early reports indicate that LAMP3 mRNA is undetectable in

Protein deglycosylation human or mouse brains (Ozaki et al., 1998; Salaun et al., 2003).

However, several genome-wide association studies (GWAS) of PD

Protein lysates were obtained using a deglycosylation lysis have recently identiﬁed the locus MCCC1/LAMP3 as a risk factor for

buffer (50 mM Tris HCl pH 7.4, 150 mM NaCl, 1% IGEPAL, 0.5% the disease (Li et al., 2013; Lill et al., 2012; Pihlstrom et al., 2013).

Deoxycholate acid, 0.1 mM EDTA, 1× Protease inhibitor cocktail). Thus, the expression of Lamp3 was investigated in a mouse model

30 ␮g of protein were used for deglycosylation using PNGase that overexpresses ␣-synuclein, a crucial protein in the pathogen-

F, which removes N-glycosylation; or with deglycosylation mix, esis of PD or in control animals with the same genetic background

which removes N- and O-glycosylation (New England Biolabs) (wild type). This animal model overexpresses the mutant human

␣

according to supplier instructions. Molecular weight shifts were A53T- -synuclein under the control of the neuron-speciﬁc prion

analyzed by western blot. promoter. The animals display normal motor activity at 6 months,

but progress to impaired spontaneous movement by 18 months,

Caspase-3 activity assay without detectable neuronal loss in the nigrostriatal projection

or protein degradation and aggregation problems (Gispert et al.,

Cells were lysed in caspase-3 lysis buffer (10 mM HEPES pH 2003). Quantitative RT-PCR analysis showed that Lamp3 was unde-

7.4, 150 mM NaCl, 1 mM EDTA, 0.5% Chaps, 1 mM Dithiothrietol, tectable in cerebellum, hippocampus, striatum, and midbrain of

1 ␮g/ml Pepstatin A, and 1 mM PMSF). Cells were sonicated and wild type mice between 3 and 20 months old. Similarly, Lamp3

J.A. Dominguez-Bautista et al. / European Journal of Cell Biology 94 (2015) 148–161 151

Table 1

Lamp3 expression in mouse brain regions. Speciﬁc brain regions were dissected, total RNA was isolated and Lamp3 expression was measured by qRT-PCR.

a a

Genotype or treatment Tissue Age or duration of treatment n Lamp3 (Ct) Tbp (Ct)

Wild type Cerebellum 3 months 5 Undetectable 27.35 ± 0.09

Hippocampus 20 months 3 Undetectable 27.76 ± 0.01

Striatum 6 months 6 Undetectable 27.10 ± 0.06

Striatum 20 months 6 Undetectable 27.52 ± 0.10

Midbrain 6 months 6 Undetectable 27.78 ± 0.07

Midbrain 20 months 6 Undetectable 28.18 ± 0.02

A53T-␣-synuclein Striatum 20 months 6 Undetectable 27.32 ± 0.09

Midbrain 6 months 6 Undetectable 27.64 0.12

Midbrain 20 months 6 Undetectable 28.00 ± 0.08

DMSO Midbrain 24 h 4 Undetectable 27.71 ± 0.12

Epoxomicin Midbrain 24 h 4 Undetectable 27.61 ± 0.06

DMSO Midbrain 2 weeks 4 Undetectable 27.42 ± 0.04

Epoxomicin Midbrain 2 weeks 4 Undetectable 27.51 ± 0.04

Mouse lung 1 25.03 32.11

Ct values of Lamp3 and Tbp were determined by qRT-PCR using mouse-speciﬁc Lamp3 TaqMan probes.

Expression of Lamp3 in mouse lung is shown for comparative purposes.

Table 2

and the cell-speciﬁc expression pattern of LAMP3 (de Saint-Vis

Ct values of LAMP genes in SH-SY5Y cells. Total RNA was isolated from cells in control

et al., 1998; Kanao et al., 2005; Ozaki et al., 1998) suggest that these

conditions and genes were detected by qRT-PCR.

proteins have distinct functions. Thus, the subcellular localization

Gene Ct values

of LAMP3 and LAMP2 was compared. A LAMP3-GFP fusion protein

TBP 26.54 ± 0.68 was generated and transiently expressed in HeLa cells; one day

LAMP1 24.03 0.10 after transfection LAMP2 was visualized by antibody staining and

LAMP2 27.23 ± 0.81

the co-localization of both proteins was determined by confocal

LAMP3 32.24 0.54

scanning microscopy. The images show that LAMP3-GFP colocal-

Ct values were obtained by qRT-PCR in at least 3 independent measurements.

izes with LAMP2 with a mean Pearson correlation coefﬁcient of

0.78. This occurs mostly in the perinuclear area, while in the cellu-

lar periphery a signiﬁcant number of vesicles are only positive for

mRNA was undetectable in striatum or midbrain of 6- and 20-

either LAMP3-GFP or LAMP2 (Fig. 1A).

month-old transgenic mice (Table 1).

A common characteristic of the LAMP family members is

Since the GWAS are based on human probands, the expression

their posttranslational glycosylation. The expected primary size

of LAMP3 was investigated in post-mortem brain tissue of PD cases

of LAMP3-GFP without posttranslational modiﬁcations is ∼70 kDa.

and age-/gender-matched controls. Braak staging indicated that PD

Western blot analysis of LAMP3-GFP lysates using an anti-GFP

cases were in stage 6 (Table S1), showing substantial Lewy body and

antibody detected two bands: the main band at ∼90 kDa, and

neurite pathology in the cingulate gyrus. Despite low RNA integrity

a weaker band at ∼130 kDa. In deglycosylated samples, LAMP3-

due to postmortem time, qRT-PCR could be successfully performed

GFP immunoreactivity appeared also in two bands, at ∼70 kDa

for the housekeeping genes EIF4A2, GAPDH, and TBP. Nevertheless,

and ∼110 kDa, indicating a deglycosylation size shift of ∼20 kDa

analysis of LAMP3 showed average Ct values of ∼37, close to a

(Fig. 1B). These results indicate that LAMP3 exists with different

non-speciﬁc signal (Fig. S-1A). Accordingly, comparing the levels

degrees of posttranslational modiﬁcation: 20 kDa due to posttrans-

of LAMP3 in PD cases versus controls showed no differences (Fig.

lational glycosylation and an additional modiﬁcation of 40 kDa of

S-1B). To determine whether the LAMP3 protein was present, west-

unknown nature. In summary, these results suggest that the LAMP

ern blot analysis was performed. PD samples and controls showed

proteins share posttranslational glycosylation, but their partially

a strong GAPDH signal, but no detectable LAMP3 protein (Fig. S-

overlapping intracellular localization (Fig. 1A) and differential tis-

1C). These data indicate that LAMP3 is not expressed in the mouse

sue expression (de Saint-Vis et al., 1998; Kanao et al., 2005; Ozaki

brain or in the human cingulate gyrus, even when PD pathology is

present. et al., 1998; Saftig et al., 2010) suggest distinct functional roles.

Tumor transformation is a factor that correlates with LAMP3

expression (Kanao et al., 2005; Ozaki et al., 1998). To determine LAMP3 is strongly induced by proteasomal inhibition in

whether LAMP3 is present in brain cancer cells, the expression neuroblastoma cells

of LAMP1, 2, and 3 was determined in human neuroblastoma SH-

SY5Y cells by qRT-PCR. Table 2 shows that LAMP1 and LAMP2 are To determine the functional role of LAMP3, several stimuli and

expressed strongly in SH-SY5Y cells, with Ct values comparable to stressors were used to determine their inﬂuence on LAMP3 expres-

those of the housekeeping gene TBP. In contrast, LAMP3 displays Ct sion in neuroblastoma SH-SY5Y cells. Measurement of transcript

values that reﬂect relatively low expression, but still signiﬁcantly levels showed that the following stimuli resulted in the downreg-

above non-speciﬁc signals (Ct values >35). This analysis indicates ulation of LAMP3: retinoic acid-induced differentiation, inhibition

that LAMP3 is present in neuroblastoma cells, probably due to its of the MEK/ERK pathway with U0126, inhibition of the PI3K/Akt

tumorigenic nature as reported for other cell lines (Kanao et al., pathway with LY294002, mitochondrial uncoupling with CCCP, as

2005; Nagelkerke et al., 2011; Pennati et al., 2013). well as induction of autophagy by the mTOR inhibitor rapamycin

or serum-/amino acid deprivation with HBSS medium. In contrast,

LAMP3 and LAMP2 colocalize in the perinuclear area but not in staurosporine did not affect LAMP3 mRNA expression signiﬁcantly.

the cellular periphery The following stressors promoted the induction of LAMP3: block-

age of autophagy with the V-ATPase inhibitor baﬁlomycin A1 or the

The different phenotypes of knockout mice lacking either PI3K inhibitor 3-methyladenine, and inhibition of the proteasome

LAMP1 (Andrejewski et al., 1999) or LAMP2 (Tanaka et al., 2000), with MG132 or epoxomicin (Fig. 2A). The upregulation of LAMP3

152 J.A. Dominguez-Bautista et al. / European Journal of Cell Biology 94 (2015) 148–161

Fig. 1. Differential localization of LAMP3 and LAMP2. (A) HeLa cells were transfected with the plasmid LAMP3-GFP (green) and one day after transfection stained with

an anti-LAMP2 antibody (red). LAMP3-GFP showed colocalization with the lysosomal protein LAMP2 as indicated by the Pearson correlation coefﬁcient especially in the

perinuclear zone, while in the cellular periphery other vesicles were only positive for LAMP3-GFP or LAMP2 (bar = 10 ␮m). Quantiﬁcation was done on three independent

slides with 5 ﬁelds of view/slide. (B) HeLa cells were transfected with LAMP3-GFP. After 24 h protein was collected in deglycosylation lysis buffer and enzymatically N- and

O-deglycosylated. Both glycosylated and deglycosylated samples were analyzed by western blot using an antibody against GFP. Two LAMP3-GFP bands were detected (∼90

and ∼130 kDa), indicating two different degrees of posttranslational modiﬁcation. Deglycosylation of LAMP3-GFP resulted in a shift of the two original bands, one of which

appeared at the predicted molecular weight of ∼70 kDa corresponding to the primary structure of the fusion protein. Another band was detected at ∼110 kDa, suggesting

the presence of a posttranslational modiﬁcation resistant to glycosylation. * indicates non-speciﬁc band.

after proteasomal inhibition was rather unusual, because MG132 Autophagic and proteasomal activity differentially regulate the

treatment commonly resulted in downregulation or had no effect expression of LAMP1, LAMP2, and LAMP3

on PD-associated genes (Fig. S-2). Thus, our data suggest that LAMP3

participates in the stress response to proteasomal inhibition. The next question was if other members of the LAMP fam-

In order to determine if these results obtained with a neuroblas- ily are equally affected by proteasomal inhibition or autophagy

toma cell line could be replicated in vivo, mouse brains were infused induction. Therefore, the expression of LAMP1, LAMP2 and LAMP3

with the proteasome inhibitor epoxomicin for 24 h or 2 weeks, was measured after starvation or proteasomal inhibition. Incuba-

and the expression of Lamp3 was measured. The inhibitory effect tion of SH-SY5Y cells in HBSS medium for 24 h had no effect on

of epoxomicin at 24 h is evidenced by the upregulation of Sqstm1 LAMP1 expression, resulted in upregulation of LAMP2 (1.6-fold), and

(sequestosome 1, synonymous to p62) to 1.3-fold (Fig. S-3), while caused downregulation of LAMP3 (0.36-fold) (Fig. 2B). The concen-

the effect of epoxomicin at 2 weeks was previously documented as tration of the proteasomal inhibitor MG132 was lowered from now

a ∼60% loss of tyrosine-hydroxylase-positive neurons in the mid- on to 300 nM for 24 h, as opposed to 3 ␮M during 16 h in Fig. 2A, to

brain compared to vehicle-infused animals (Subramaniam et al., diminish cell death. Proteasomal inhibition had no effect on LAMP1

2014). As shown in Table 1, Lamp3 expression was undetectable or LAMP2, and induced a strong upregulation of LAMP3 (10.3-fold)

after 24 h and 2 weeks of treatment, suggesting that additional fac- (Fig. 2C). Importantly, the strong effect of proteasomal inhibition

tors as e.g. tumorigenic transformation are needed to induce Lamp3 on LAMP3 expression could be reproduced in an additional cell line:

expression in the brain. proteasomal inhibition in HeLa cells resulted also in upregulation

J.A. Dominguez-Bautista et al. / European Journal of Cell Biology 94 (2015) 148–161 153

Fig. 2. LAMP3 mRNA levels were modulated by starvation-induced autophagy and proteasomal inhibition. (A) Relative levels of LAMP3 mRNA were determined by qRT-PCR in

SH-SY5Y cells under different conditions. Cellular differentiation was induced with 10 ␮M retinoic acid (RA) and 3% FCS for 10 days. The following treatments had a duration

of 16 h: control medium with 0.05% DMSO, 10 ␮M U0126, 25 ␮M LY294002 (LY), 10 nM CCCP, 0.5 ␮M rapamycin (Rap), starvation medium (HBSS), 30 nM staurosporine

(Stau), 10 nM baﬁlomycin A (Baf), 10 mM 3-methyladenine (3MA), 3 ␮M MG132, and 1 ␮M epoxomicin (Epo). LAMP3 was downregulated by RA-induced differentiation,

U0126, LY294002, CCCP, rapamycin- and starvation-induced autophagy. Staurosporine did not affect LAMP3. Autophagy inhibition with baﬁlomycin A1 or 3-methyladenine,

and proteasome inhibition with MG132 or epoxomicin induced LAMP3 upregulation (n = 3). (B) SH-SY5Y cells were cultured in complete (RPMI + 10%FCS) or in starvation

medium (HBSS) for 24 h. Relative mRNA levels of LAMP1, LAMP2, and LAMP3 were determined by qRT-PCR. LAMP1 mRNA levels remained unchanged, LAMP2 mRNA levels

were signiﬁcantly upregulated, and LAMP3 mRNA levels were signiﬁcantly downregulated by starvation compared to the control (n = 3). (C) SH-SY5Y cells were cultured

in the presence of 300 nM MG132 or the equivalent concentration of the vehicle (DMSO) for 24 h. Relative mRNA levels of LAMP1, LAMP2, and LAMP3 were determined by

qRT-PCR. LAMP1 and LAMP2 mRNA levels remained unchanged, and LAMP3 mRNA was robustly induced by proteasomal inhibition compared to the vehicle-treated control

(n = 4–7). Statistically signiﬁcant values of t-tests are indicated by *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001.

of LAMP3 (3.6-fold), but not of LAMP1 (0.7-fold) or of LAMP2 (0.6- control siRNA and then exposed to MG132 or DMSO as control for

fold) (Fig. S-4). These results imply that in response to proteasomal 24 h. Quantitative RT-PCR analyses showed that MG132 treatment

dysfunction LAMP3 is speciﬁcally upregulated. induced the expression of ATF4 (2.1-fold), LAMP3 (10-fold), as well

Next, protein expression of LAMP3 was analyzed. The LAMP3 as the UPR genes HSPA5 (4.6-fold), and ATF3 (33-fold) (Fig. 3A). The

∼

AF4087 antibody recognized bands of the expected size in the pos- 80% reduction of ATF4 in MG132-treated cells correlated with a

itive controls (human lung tissue and HeLa cells overexpressing reduction of LAMP3, HSPA5, and ATF3 expression by 90%, 40%, and

LAMP3-GFP), but both in SH-SY5Y and HeLa cells a strong non- 60%, respectively. In contrast, LAMP1 expression was not affected by

speciﬁc signal covered the region of the endogenous LAMP3 band ATF4 knockdown, and although LAMP2 was not induced by MG132

(Fig. S-5A). Usage of other commercial antibodies (Fig. S-5B and C) alone, its expression was partially inhibited by ATF4 siRNA (∼0.6-

or custom-generated LAMP3 antibodies resulted in nonspeciﬁc sig- fold) (Fig. 3B). These data indicate that during proteasomal stress

nals. Thus, it was not possible to conﬁrm the transcriptional data LAMP3 is induced as part of the UPR in an ATF4-dependent pathway.

at the protein level. In the reverse experiment the possibility that the knockdown of

LAMP3 exacerbates the UPR and feeds-back onto UPR gene regula-

LAMP3 induction during proteasomal inhibition depends on the tion was investigated. To this end, SH-SY5Y cells were transfected

ATF4 transcription factor with control (CTRL siRNA) or LAMP3-targeted siRNA (LAMP3 siRNA)

before initiating MG132 treatment. Afterwards, gene expression of

Previous data showed that LAMP3 induction by different stress- the UPR genes HSPA5, DDIT3, ATF3 and ATF4 was determined by qRT-

ors is depending on the UPR transcription factor ATF4 (Mujcic PCR. Basal levels of LAMP3 were reduced by 50%, and in the presence

et al., 2009; Nagelkerke et al., 2013a,b). To determine whether the of MG132 the 10-fold upregulation was blunted to 2.3-fold (Fig. 4A).

ATF4 transcription factor controls the expression of LAMP3 during Only in the case of HSPA5 the silencing of LAMP3 resulted in

proteasomal stress, the expression of molecules involved in UPR further upregulation of the corresponding transcript, thus show-

and LAMP genes was measured in cells with and without ATF4 ing that lack of LAMP3 is not enough to exacerbate generalized

knockdown. SH-SY5Y cells were transfected with ATF4 siRNA or UPR gene expression (Fig. 4B). In contrast to the MG132-induced

154 J.A. Dominguez-Bautista et al. / European Journal of Cell Biology 94 (2015) 148–161

Fig. 3. LAMP3 induction during proteasomal inhibition depends on the transcription factor ATF4. (A) SH-SY5Y cells were transfected either with control siRNA (CTRL siRNA)

or with ATF4-directed siRNA (ATF4 siRNA) and after 1 d treated for 24 h with 300 nM MG132 or vehicle. Transcript levels of ATF4, LAMP3, HSPA5, and ATF3 were determined

by qRT-PCR. ATF4 was induced by MG132 and its induction was inhibited with ∼80% efﬁciency in ATF4 siRNA transfected samples. ATF4 knockdown prevented the induction

of LAMP3, and the UPR genes HSPA5 and ATF3 (n = 3). (B) Using the same experimental set-up as in (A), transcript levels of LAMP1 and LAMP2 were measured. LAMP1 was

not affected by ATF4 knockdown. LAMP2 was not induced due to proteasomal inhibition, but its expression was inhibited by ATF4 siRNA during MG132 treatment (n = 3).

Statistical analysis was performed by ANOVA with Bonferroni’s post hoc test *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001.

upregulation of the UPR genes, analysis of lysosomal and thapsigargin or tunicamycin, known ER stressors and activators

autophagic genes showed either downregulation or no effect in of the UPR. Both thapsigargin and tunicamycin treatment resulted

response to MG132 treatment, and this was independent of LAMP3 indeed in a signiﬁcant upregulation of LAMP3 and the UPR molecule

(Table S2). HSPA5, but not LAMP1 or LAMP2 (Fig. S-6). Taken together, these

To substantiate our hypothesis that the UPR is involved in ﬁndings support the notion that LAMP3 is induced as part of the

the induction of LAMP3, SH-SY5Y cells were incubated with UPR in SH-SY5Y cells during proteasomal inhibition.

J.A. Dominguez-Bautista et al. / European Journal of Cell Biology 94 (2015) 148–161 155

Fig. 4. LAMP3 regulates HSPA5 gene expression during proteasomal inhibition. (A) SH-SY5Y cells were transfected with control siRNA (CTRL siRNA) or with LAMP3-targeted

siRNA (LAMP3 siRNA) and after 1 d cells were treated for 24 h with 300 nM MG132 or equivalent concentration of the vehicle DMSO. Relative levels of LAMP3 were determined

by qRT-PCR. Basal LAMP3 levels and LAMP3 upregulation induced by MG132 were signiﬁcantly reduced by LAMP3 siRNA treatment (n = 8). (B) Using the same experimental

set-up as in (A), relative mRNA levels of HSPA5, DDIT3, ATF3, and ATF4 were determined by qRT-PCR in SH-SY5Y cells. MG132-mediated proteasomal inhibition resulted in

upregulation of HSPA5, DDIT3, ATF3, and ATF4. Prevention of LAMP3 upregulation resulted in an increase of HSPA5, but had no effect on the other genes (n = 3). Statistical

analysis was performed by ANOVA with Bonferroni’s post hoc test *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001.

LAMP3 upregulation is contributing to autophagic ﬂux induction proteasomal inhibition was assessed by measuring SQSTM1/p62 at

after proteasomal inhibition the transcript and protein level, as well as the formation of LC3-II

as parameters for autophagy and autophagic ﬂux (Mizushima et al.,

Proteasomal dysfunction prompts compensatory autophagy 2010). SQSTM1was determined in SH-SY5Y cells with and without

(Korolchuk et al., 2010; Nedelsky et al., 2008; Park and Cuervo, LAMP3 knockdown and MG132 exposure. Measurement of SQSTM1

2013), the unfolded protein response (UPR) being one of the pro- mRNA showed an 11-fold induction upon MG132 treatment, and

posed crosstalk pathways between these two systems (Korolchuk this induction was not affected by knockdown of LAMP3 (Fig. 5A).

et al., 2010). Therefore, the inﬂuence of LAMP3 on autophagy during At the protein level, SQSTM1 was expressed only minimally in

156 J.A. Dominguez-Bautista et al. / European Journal of Cell Biology 94 (2015) 148–161

Fig. 5. SQSTM1 induction and puncta formation is independent of LAMP3. (A) SH-SY5Y cells were transfected with CTRL siRNA or LAMP3 siRNA and after 1 d exposed to 300 nM

MG132 or DMSO for 24 h. SQSTM1 was measured by qRT-PCR. Proteasomal inhibition with MG132 resulted in upregulation of SQSTM1 (n = 3). (B) Cells were transfected and

treated with MG132 as in (A) and total protein was analyzed by western blot using an antibody against SQSTM1. A representative western blot as well as the densitometric

analysis is shown. SQSTM1 was strongly induced by MG132 treatment, and the extent of its induction was not affected by silencing of LAMP3 (n = 3). (C) Cells were transfected

and treated with MG132 as in (A), and stained with an antibody against SQSTM1. The number of SQSTM1 puncta per cell was determined by confocal microscopy. The number

of puncta per cell was increased by MG132 treatment, independently of LAMP3 induction. A total of 20–70 cells in at least 5 randomly chosen ﬁelds of view were used in the

quantiﬁcation. A representative image is shown in Fig. S-7A. Statistical analysis was performed by ANOVA with Bonferroni’s post hoc test *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001.

J.A. Dominguez-Bautista et al. / European Journal of Cell Biology 94 (2015) 148–161 157

control conditions; upon MG132 treatment its expression In summary, our data show that upregulation of LAMP3 in SH-

increased strongly (∼50-fold) and was not inﬂuenced by the silenc- SY5Y cells is necessary for the response to proteasomal inhibition

ing of LAMP3 (Fig. 5B). Treatment with MG132 increased the in order to promote cell survival.

number of SQSTM1 signals (puncta), which are most probably asso-

ciated with LC3-positive autophagosomes (Sahani et al., 2014).

However, their number was not affected by the knockdown of Discussion

LAMP3 (Fig. 5C and Fig. S-7A).

Another crucial marker to monitor autophagy is the assess- Recent genome-wide association studies of PD have identiﬁed

ment of LC3-I conversion to LC3-II in the presence and absence of variants at the locus MCCC1/LAMP3 as risk factors, although LAMP3

inhibitors of autophagosome-lysosome fusion, as baﬁlomycin A1 was not detected in brain tissue (Akasaki et al., 2004; de Saint-

(Mizushima et al., 2010). We showed before that in SH-SY5Y cells a Vis et al., 1998; Kanao et al., 2005; Salaun et al., 2004). With this

LC3-II band in the absence of baﬁlomycin A1 is not detectable even ﬁnding as starting point we investigated a possible function that

in cells starved for more than 12 h (Klinkenberg et al., 2012), prob- LAMP3 could have in the progression of PD. Neither overexpres-

ably due to strong autophagic ﬂux. Therefore, analysis of LC3-II in sion of the mutant human A53T-␣-synuclein nor the infusion of a

SH-SY5Y cells was monitored only in the presence of baﬁlomycin proteasomal inhibitor induced the expression of LAMP3 in mouse

A1. As predicted, proteasomal inhibition resulted in an upregula- brain. Furthermore, LAMP3 mRNA and protein was not detected in

tion of autophagic ﬂux (ratio LC3-II/GAPDH) (3.1-fold). Importantly, the posterior cingulate gyrus of PD cases and age-/gender-matched

siRNA-mediated LAMP3 silencing resulted in an impaired induc- controls. Recently, Murphy and colleagues reported that early and

tion of autophagic ﬂux (1.9-fold) (Fig. 6A), indicating that LAMP3 late cases of PD as well as age-matched controls expressed LAMP3

is necessary for the induction of autophagy upon proteasomal dys- protein in the anterior cingulate and that there were no signiﬁ-

function despite unchanged SQSTM1 levels (see Discussion). cant differences between these two groups (Murphy et al., 2014).

As LAMP3 is a lysosomal protein, potential changes in lysosome In our hands, the same antibody also detected a band at ∼65 kDa

number and distribution after LAMP3 knockdown were assessed. in HeLa cells. But this signal was not affected by LAMP3 siRNA

Cells were transfected with LAMP3 siRNA or control siRNA, treated transfection or deglycosylation; thus, we assume that it represents

with MG132 or DMSO, and immunostained for the lysosomal a non-speciﬁc band (Fig. S-5B). While the expression of LAMP3

marker LAMP1. No signiﬁcant changes were observed in the dis- in other brain regions should not be ruled out, our data indicate

tribution or the amount of LAMP1-positive vesicles per cell (Fig. 6B no direct association of LAMP3 with the neuronal pathomecha-

and Fig. S-7B). This suggests that the alteration in autophagic ﬂux nisms that are contributing to PD. Further investigation is needed

during MG132 treatment is not mediated by changes in lysosomal to understand how the MCCC1/LAMP3 locus contributes to the initi-

number. ation or progression of PD. Possible mechanisms could include the

To determine whether the effect of LAMP3 siRNA on autophagic participation of LAMP3 in cellular physiology outside the brain, or

flux is specific for proteasomal inhibition, the influence of LAMP3 the involvement of other genes present in the MCCC1/LAMP3 locus.

on starvation-induced autophagy was measured. Fig. 6C shows In contrast to the absence of LAMP3 in brain, the cell line SH-

that the autophagic ﬂux in cells cultured in starvation medium SY5Y, originating from a neuroblastoma tumor, expresses this gene.

was not affected by the knockdown of LAMP3. This indicates that The LAMP3 expression is probably due to the tumorigenic nature

LAMP3 specifically influences autophagic flux induced by proteaso- of the cell line, as LAMP3 is involved in cellular transformation and

mal inhibition, but not by starvation. invasiveness in cervical, breast, and prostate cancer cells (Kanao

et al., 2005; Nagelkerke et al., 2011; Pennati et al., 2013). In support

LAMP3 is necessary for cell survival during proteasomal inhibition of this hypothesis, differentiation of SH-SH5Y cells with retinoic

acid reduced LAMP3 levels by 50%.

As shown above and detailed in the discussion, autophagy Using different pharmacologic modulators, proteasomal

induction is a fundamental cellular survival response to protea- inhibitors were identiﬁed as strong inducers of LAMP3 expres-

somal dysfunction. Given that autophagy induction as a result of sion. Previous reports have shown that the LAMP3 expression is

proteasomal inhibition was impaired in LAMP3-silenced cells, the ATF4-dependent during hypoxia (Nagelkerke et al., 2013a,b). Our

role of LAMP3 in cell survival was evaluated. SH-SY5Y cells were results imply that the mechanism that controls LAMP3 expression

transiently transfected with the LAMP3-targeted siRNA or a control during proteasomal stress is the UPR because (1) knockdown

siRNA, and either treated with MG132 or DMSO. After 24 h apo- of the UPR transcription factor ATF4 prevented the induction

ptosis was quantiﬁed by two methods that determine caspase-3 of LAMP3 by MG132 treatment, (2) other molecules of the UPR

activation. Proteasomal inhibition resulted in a strong induction of were induced similar to LAMP3, and (3) direct activation of the

apoptosis measured by caspase-3 activity (13.5-fold), which was UPR with thapsigargin or tunicamycin strongly triggered LAMP3,

further increased in cells that were previously transfected with the but not LAMP1 or LAMP2 induction. Thus, our data indicate that

siRNA targeting LAMP3 (17.8-fold) (Fig. 7A). This result could be proteasomal inhibition activates LAMP3 expression via the UPR

conﬁrmed by measuring the cleavage of PARP, an early substrate of in neuroblastoma cells. Despite a lot of effort dedicated to this

caspase-3. Proteasomal inhibition resulted in increased appearance subject, the induction of LAMP3 at the protein level could not

of the 89 kDa PARP fragment, and the prevention of LAMP3 upreg- be determined due to crossreactivity or nonspeciﬁcity of several

ulation resulted in a further increase of PARP cleavage (Fig. 7B). commercial LAMP3 antibodies.

A rescue experiment by transfecting SH-SY5Y cells with LAMP3- Consistent with other reports (Lan et al., 2015; Tang et al., 2014),

GFP prior to MG132 treatment was not successful, due to the low analysis of the autophagosome marker LC3-II showed that protea-

transfection efﬁciency of LAMP3-GFP in this cell line. somal inhibition promoted enhancement of the autophagic ﬂux.

To determine whether the effect of LAMP3 knockdown on apo- Silencing of LAMP3 impaired the induction of autophagic ﬂux and

ptosis is speciﬁc for proteasomal inhibition, cells were transfected promoted cell death. As indicator of autophagic ﬂux we relied

with LAMP3 siRNA and treated for 6 h with the kinase inhibitor stau- on the autophagosomal marker LC3-II rather than SQSTM1/p62.

rosporine, a well-known inducer of apoptosis. As shown in Fig. 7C, SQSTM1 interacts on the one hand with the ubiquitin moiety in

exposure to staurosporine resulted in a strong increase of caspase- ubiquitin-tagged protein substrates, and on the other hand with

3 activity, which was however not inﬂuenced by LAMP3. These data LC3-II for subsequent autophagic degradation (Lamark et al., 2009).

point to a speciﬁc function of LAMP3 in proteasomal inhibition. As can be seen in Fig. 5B and C, the basal levels of SQSTM1 protein

158 J.A. Dominguez-Bautista et al. / European Journal of Cell Biology 94 (2015) 148–161

Fig. 6. Prevention of LAMP3 mRNA upregulation resulted in impaired autophagic ﬂux during proteasomal inhibition. (A) Autophagic ﬂux was analyzed in SH-SY5Y cells with

or without LAMP3 knockdown either in control conditions (DMSO) or in the presence of 300 nM MG132 for 24 h. Baﬁlomycin A1 was added at a ﬁnal concentration of 10 nM

during the last 2 h of culture. Total protein was analyzed by western blot using an anti-LC3 antibody, and GAPDH as loading control. A representative western blot as well as

the densitometric analysis is shown. Autophagic ﬂux was impaired by LAMP3 siRNA in MG132-treated cells (n = 3). (B) The number of LAMP1-positive vesicles per cell was

quantiﬁed in cells with or without LAMP3 knockdown and treated with MG132 or DMSO. Cells were ﬁxed, stained with an antibody against LAMP1, and analyzed by confocal

microscopy. The number of LAMP1-positive vesicles was not affected by MG132 treatment or LAMP3 knockdown. A total of 20–70 cells in at least 5 randomly chosen ﬁelds

of view were used for the quantiﬁcation. A representative image is shown in Fig. S-7B. (C) Autophagic ﬂux was analyzed in SH-SY5Y cells with or without LAMP3 knockdown

either in control conditions (RPMI) or in starvation medium (HBSS) for 2 h. In all cases 10 nM baﬁlomycin A1 was present in the culture. Total protein was analyzed by

western blot using an anti-LC3 antibody, and GAPDH as loading control. A representative western blot as well as the densitometric analysis is shown. Autophagic ﬂux was

not inﬂuenced by LAMP3 siRNA in starved cells (n = 4). Statistical analysis was performed by ANOVA with Bonferroni’s post hoc test *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001.

J.A. Dominguez-Bautista et al. / European Journal of Cell Biology 94 (2015) 148–161 159

Fig. 7. Prevention of LAMP3 mRNA upregulation resulted in increased apoptosis during proteasomal inhibition. (A) SH-SY5Y cells were transfected with CTRL siRNA or LAMP3

siRNA and after 1 d exposed to 300 nM MG132 or DMSO for 24 h. Caspase-3 activity analysis was performed as indicated in materials and methods. Knockdown of LAMP3

resulted in an increase of MG132-induced caspase-3 activity. Results were normalized to protein amount and shown relative to DMSO + CTRL siRNA (n = 5). (B) SH-SY5Y cells

were transfected with CTRL siRNA or LAMP3 siRNA, cultured overnight and then exposed to 300 nM MG132 or DMSO for 24 h. Cleavage of Poly-ADP Ribose Polymerase (PARP)

was determined by western blot. Increased cleaved PARP 89 kDa was observed upon proteasomal inhibition, and a further increase was observed in LAMP3 siRNA-transfected

cells (n = 4). (C) SH-SY5Y cells were transfected with CTRL siRNA or LAMP3 siRNA and after 1 d exposed to 30 nM staurosporine or DMSO for 6 h. Caspase-3 activity analysis

was performed as indicated in materials and methods. Knockdown of LAMP3 did not affect caspase-3 activity induced by staurosporine. Results were normalized to protein

amount and shown relative to DMSO + CTRL siRNA (n = 3–5). Statistical analysis was performed by ANOVA with Bonferroni’s post hoc test *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001.

160 J.A. Dominguez-Bautista et al. / European Journal of Cell Biology 94 (2015) 148–161

were very low and increased strongly after MG132-triggered pro- network (BMBF 01EW1012), and by the GerontoMitoSys network

teasomal inhibition. It seems likely that an additional effect due to (BMBF 0315584A).

LAMP3 silencing would not be detectable in this strong reaction.

The mechanism that mediates the induction of autophagic ﬂux

Appendix A. Supplementary data

following proteasome inhibition seems to depend on the UPR. In

support of this hypothesis, it was reported that the transcription

Supplementary data associated with this article can be found, in

factor ATF4 mediates autophagy in response to the proteasome

the online version, at http://dx.doi.org/10.1016/j.ejcb.2015.01.003.

inhibitors bortezomib and oprozomib in multiple myeloma and

other cancers (Rzymski et al., 2009; Zang et al., 2012). Therefore,

we propose the following scenario in our model: proteasomal inhi- References

bition with MG132 activates the ATF4 pathway of the UPR, thus

driving LAMP3 expression and promoting autophagic ﬂux. During Akasaki, K., Nakamura, N., Tsukui, N., Yokota, S., Murata, S., Katoh, R., Michihara, A.,

Tsuji, H., Marques Jr., E.T., August, J.T., 2004. Human dendritic cell lysosome-

proteasomal inhibition enhancement of the autophagic pathway

associated membrane protein expressed in lung type II pneumocytes. Arch.

acts as compensatory response to degrade accumulated protea-

Biochem. Biophys. 425, 147–157.

some substrates. It is important to stress MG132-mediated ATF4 Andrejewski, N., Punnonen, E.L., Guhde, G., Tanaka, Y., Lullmann-Rauch, R., Hart-

mann, D., von Figura, K., Saftig, P., 1999. Normal lysosomal morphology and

signaling pathway is primarily targeting LAMP3 and not LAMP1,

function in LAMP-1-deﬁcient mice. J. Biol. Chem. 274, 12692–12701.

and only slightly inﬂuencing LAMP2. Given that LAMP3 and LAMP2

Braak, H., Braak, E., 1991. Neuropathological stageing of Alzheimer-related changes.

signals did not entirely colocalize in immunoﬂuorescence and Acta Neuropathol. 82, 239–259.

Braak, H., Del Tredici, K., Rub, U., de Vos, R.A., Jansen Steur, E.N., Braak, E., 2003.

that the number, appearance and distribution of lysosomes and

Staging of brain pathology related to sporadic Parkinson’s disease. Neurobiol.

p62-positive puncta/autophagosomes did not change after LAMP3

Aging 24, 197–211.

knockdown, it is tempting to speculate that LAMP3 has a role Combaret, V., Boyault, S., Iacono, I., Brejon, S., Rousseau, R., Puisieux, A., 2008. Effect

of bortezomib on human neuroblastoma: analysis of molecular mechanisms

in autophagosome maturation and/or autophagosome-lysosome

involved in cytotoxicity. Mol. Cancer 7, 50.

fusion.

de Saint-Vis, B., Vincent, J., Vandenabeele, S., Vanbervliet, B., Pin, J.J., Ait-Yahia,

In case of LAMP3 deﬁciency, autophagy induction is impaired, S., Patel, S., Mattei, M.G., Banchereau, J., Zurawski, S., Davoust, J., Caux, C.,

Lebecque, S., 1998. A novel lysosome-associated membrane glycoprotein, DC-

and the autophagy-lysosomal pathway is overtaxed resulting in

LAMP, induced upon DC maturation, is transiently expressed in MHC class II

further accumulation of non-degraded substrates. With both of the

compartment. Immunity 9, 325–336.

major degradation pathways dysfunctional, apoptosis is enhanced. Erraﬁy, R., Aguado, C., Ghislat, G., Esteve, J.M., Gil, A., Loutﬁ, M., Knecht, E., 2013. PTEN

This hypothesis is supported by previous studies revealing the increases autophagy and inhibits the ubiquitin-proteasome pathway in glioma

cells independently of its lipid phosphatase activity. PLOS ONE 8, e83318.

participation of LAMP3 in autophagy and cellular viability in dif-

Eskelinen, E.L., Schmidt, C.K., Neu, S., Willenborg, M., Fuertes, G., Salvador, N., Tanaka,

ferent cancer cells. One report showed that breast cancer cells

Y., Lullmann-Rauch, R., Hartmann, D., Heeren, J., von Figura, K., Knecht, E., Saftig,

resistant to the cytostatic drug tamoxifen expressed 7-times more P., 2004. Disturbed cholesterol trafﬁc but normal proteolytic function in LAMP-

1/LAMP-2 double-deﬁcient ﬁbroblasts. Mol. Biol. Cell 15, 3132–3145.

LAMP3 mRNA than sensitive cells, and exhibited signiﬁcantly higher

Furfaro, A.L., Piras, S., Passalacqua, M., Domenicotti, C., Parodi, A., Fenoglio, D.,

autophagic activity. Importantly, siRNA-mediated LAMP3 silencing

Pronzato, M.A., Marinari, U.M., Moretta, L., Traverso, N., Nitti, M., 2014. HO-1

resulted in autophagy inhibition and restored tamoxifen sensitivity up-regulation: a key point in high-risk neuroblastoma resistance to bortezomib.

Biochim. Biophys. Acta 1842, 613–622.

(Nagelkerke et al., 2014). Furthermore, prostate cancer cells resis-

Gispert, S., Del Turco, D., Garrett, L., Chen, A., Bernard, D.J., Hamm-Clement, J., Korf,

tant to the chemotherapeutic cisplatin can be resensitized to the

H.W., Deller, T., Braak, H., Auburger, G., Nussbaum, R.L., 2003. Transgenic mice

drug by the knockdown of LAMP3. This effect seems mediated by expressing mutant A53T human alpha-synuclein show neuronal dysfunction in

the absence of aggregate formation. Mol. Cell. Neurosci. 24, 419–429.

the impairment of autophagic ﬂux (Pennati et al., 2013).

Glickman, M.H., Ciechanover, A., 2002. The ubiquitin-proteasome proteolytic path-

As recent data indicate that autophagy and the ubiquitin-

way: destruction for the sake of construction. Physiol. Rev. 82, 373–428.

proteasome system have crucial roles for tumor development and He, C., Klionsky, D.J., 2009. Regulation mechanisms and signaling pathways of

autophagy. Annu. Rev. Genet. 43, 67–93.

metastasis (Erraﬁy et al., 2013; Qiang et al., 2014; Vogl et al., 2014),

Kanao, H., Enomoto, T., Kimura, T., Fujita, M., Nakashima, R., Ueda, Y., Ueno, Y., Miy-

our ﬁndings that LAMP3 participates in the response to proteaso-

atake, T., Yoshizaki, T., Buzard, G.S., Tanigami, A., Yoshino, K., Murata, Y., 2005.

mal dysfunction are relevant for cancer research. The proteasomal Overexpression of LAMP3/TSC403/DC-LAMP promotes metastasis in uterine

inhibitor bortezomib is approved for the treatment of multiple cervical cancer. Cancer Res. 65, 8640–8645.

Klinkenberg, M., Gispert, S., Dominguez-Bautista, J.A., Braun, I., Auburger, G., Jen-

myeloma, but its effectiveness is limited against breast tumors

drach, M., 2012. Restriction of trophic factors and nutrients induces PARKIN

(Combaret et al., 2008; Mujtaba and Dou, 2011). In addition, cancer

expression. Neurogenetics 13, 9–21.

cells often develop resistance to chemotherapy (Furfaro et al., 2014) Korolchuk, V.I., Menzies, F.M., Rubinsztein, D.C., 2010. Mechanisms of cross-talk

between the ubiquitin-proteasome and autophagy-lysosome systems. FEBS Lett.

and the resistance mechanisms include the induction of the UPR

584, 1393–1398.

and autophagy (Milani et al., 2009). Accordingly, LAMP3 could be

Lamark, T., Kirkin, V., Dikic, I., Johansen, T., 2009. NBR1 and p62 as cargo receptors

involved in the development of resistance to proteasomal inhibitors for selective autophagy of ubiquitinated targets. Cell Cycle 8, 1986–1990.

Lan, D., Wang, W., Zhuang, J., Zhao, Z., 2015. Proteasome inhibitor-induced

during chemotherapy.

autophagy in PC12 cells overexpressing A53T mutant alpha-synuclein. Mol.

In conclusion, we report a previously unknown transcriptional

Med. Rep. 11, 1655–1660.

upregulation of LAMP3 expression after proteasomal stress, which Li, N.N., Tan, E.K., Chang, X.L., Mao, X.Y., Zhao, D.M., Zhang, J.H., Liao, Q., Peng, R., 2013.

MCCC1/LAMP3 reduces risk of sporadic Parkinson’s disease in Han Chinese. Acta

is contributing to promote autophagy and cell survival.

Neurol. Scand. 128, 136–139.

Lill, C.M., Roehr, J.T., McQueen, M.B., Kavvoura, F.K., Bagade, S., Schjeide, B.M.,

Schjeide, L.M., Meissner, E., Zauft, U., Allen, N.C., Liu, T., Schilling, M., Ander-

Acknowledgements son, K.J., Beecham, G., Berg, D., Biernacka, J.M., Brice, A., DeStefano, A.L., Do, C.B.,

Eriksson, N., Factor, S.A., Farrer, M.J., Foroud, T., Gasser, T., Hamza, T., Hardy, J.A.,

Heutink, P., Hill-Burns, E.M., Klein, C., Latourelle, J.C., Maraganore, D.M., Martin,

Brain tissue was obtained from the Neurobiobank Munich/ E.R., Martinez, M., Myers, R.H., Nalls, M.A., Pankratz, N., Payami, H., Satake, W.,

Brain-Net Germany (http://www.brain-net.net), and we are grate- Scott, W.K., Sharma, M., Singleton, A.B., Stefansson, K., Toda, T., Tung, J.Y., Vance,

J., Wood, N.W., Zabetian, C.P., Young, P., Tanzi, R.E., Khoury, M.J., Zipp, F., Lehrach,

ful to Dr. Thomas Arzberger and colleagues for it. We thank

H., Ioannidis, J.P., Bertram, L., 2012. Comprehensive research synopsis and sys-

Prof. Deller, Dr. Domenico del Turco and Heike Korff for the

tematic meta-analyses in Parkinson’s disease genetics: the PDGene database.

RNA integrity analysis of human brain tissues. We are grate- PLoS Genet. 8, e1002548.

Milani, M., Rzymski, T., Mellor, H.R., Pike, L., Bottini, A., Generali, D., Harris, A.L., 2009.

ful to Birgitt Meseck-Selchow for technical assistance. The study

The role of ATF4 stabilization and autophagy in resistance of breast cancer cells

was ﬁnancially supported by the NGFNplus Parkinson net-

treated with Bortezomib. Cancer Res. 69, 4415–4423.

work (BMBF 01GS08138), by the ERANET-NEURON-RePARK 2009 Mizushima, N., 2007. Autophagy: process and function. Genes Dev. 21, 2861–2873.

J.A. Dominguez-Bautista et al. / European Journal of Cell Biology 94 (2015) 148–161 161

Mizushima, N., Yoshimori, T., Levine, B., 2010. Methods in mammalian autophagy Qiang, L., Zhao, B., Ming, M., Wang, N., He, T.C., Hwang, S., Thorburn, A., He, Y.Y.,

research. Cell 140, 313–326. 2014. Regulation of cell proliferation and migration by p62 through stabilization

Mujcic, H., Rzymski, T., Rouschop, K.M., Koritzinsky, M., Milani, M., Harris, A.L., of Twist1. Proc. Natl. Acad. Sci. U. S. A. 111, 9241–9246.

Wouters, B.G., 2009. Hypoxic activation of the unfolded protein response (UPR) Rzymski, T., Milani, M., Singleton, D.C., Harris, A.L., 2009. Role of ATF4 in reg-

induces expression of the metastasis-associated gene LAMP3. Radiother. Oncol. ulation of autophagy and resistance to drugs and hypoxia. Cell Cycle 8,

92, 450–459. 3838–3847.

Mujtaba, T., Dou, Q.P., 2011. Advances in the understanding of mechanisms and Saftig, P., Klumperman, J., 2009. Lysosome biogenesis and lysosomal membrane

therapeutic use of bortezomib. Discov. Med. 12, 471–480. proteins: trafﬁcking meets function. Nat. Rev. Mol. Cell Biol. 10, 623–635.

Murphy, K.E., Gysbers, A.M., Abbott, S.K., Tayebi, N., Kim, W.S., Sidransky, E., Cooper, Saftig, P., Schroder, B., Blanz, J., 2010. Lysosomal membrane proteins: life between

A., Garner, B., Halliday, G.M., 2014. Reduced glucocerebrosidase is associated acid and neutral conditions. Biochem. Soc. Trans. 38, 1420–1423.

with increased alpha-synuclein in sporadic Parkinson’s disease. Brain 137, Sahani, M.H., Itakura, E., Mizushima, N., 2014. Expression of the autophagy sub-

834–848. strate SQSTM1/p62 is restored during prolonged starvation depending on

Nagelkerke, A., Bussink, J., Mujcic, H., Wouters, B.G., Lehmann, S., Sweep, F.C., transcriptional upregulation and autophagy-derived amino acids. Autophagy

Span, P.N., 2013a. Hypoxia stimulates migration of breast cancer cells via the 10, 431–441.

PERK/ATF4/LAMP3-arm of the unfolded protein response. Breast Cancer Res. Salaun, B., de Saint-Vis, B., Clair-Moninot, V., Pin, J.J., Barthelemy-Dubois, C.,

15, R2. Kissenpfennig, A., Peronne, C., Bates, E., Mattei, M.G., Lebecque, S., 2003. Cloning

Nagelkerke, A., Bussink, J., van der Kogel, A.J., Sweep, F.C., Span, P.N., 2013b. The and characterization of the mouse homologue of the human dendritic cell mat-

PERK/ATF4/LAMP3-arm of the unfolded protein response affects radioresistance uration marker CD208/DC-LAMP. Eur. J. Immunol. 33, 2619–2629.

by interfering with the DNA damage response. Radiother. Oncol. 108, 415–421. Salaun, B., de Saint-Vis, B., Pacheco, N., Pacheco, Y., Riesler, A., Isaac, S., Leroux,

Nagelkerke, A., Mujcic, H., Bussink, J., Wouters, B.G., van Laarhoven, H.W., Sweep, F.C., C., Clair-Moninot, V., Pin, J.J., Grifﬁth, J., Treilleux, I., Goddard, S., Davoust, J.,

Span, P.N., 2011. Hypoxic regulation and prognostic value of LAMP3 expression Kleijmeer, M., Lebecque, S., 2004. CD208/dendritic cell-lysosomal associated

in breast cancer. Cancer 117, 3670–3681. membrane protein is a marker of normal and transformed type II pneumocytes.

Nagelkerke, A., Sieuwerts, A.M., Bussink, J., Sweep, F.C., Look, M.P., Foekens, J.A., Am. J. Pathol. 164, 861–871.

Martens, J.W., Span, P.N., 2014. LAMP3 is involved in tamoxifen resistance in Subramaniam, M., Kern, B., Vogel, S., Klose, V., Schneider, G., Roeper, J., 2014. Selec-

breast cancer cells through the modulation of autophagy. Endocr. Relat. Cancer tive increase of in vivo ﬁring frequencies in DA SN neurons after proteasome

21, 101–112. inhibition in the ventral midbrain. Eur J Neurosci.

Nedelsky, N.B., Todd, P.K., Taylor, J.P., 2008. Autophagy and the ubiquitin- Tanaka, Y., Guhde, G., Suter, A., Eskelinen, E.L., Hartmann, D., Lullmann-Rauch, R.,

proteasome system: collaborators in neuroprotection. Biochim. Biophys. Acta Janssen, P.M., Blanz, J., von Figura, K., Saftig, P., 2000. Accumulation of autophagic

1782, 691–699. vacuoles and cardiomyopathy in LAMP-2-deﬁcient mice. Nature 406, 902–906.

Ozaki, K., Nagata, M., Suzuki, M., Fujiwara, T., Ueda, K., Miyoshi, Y., Takahashi, E., Tang, B., Cai, J., Sun, L., Li, Y., Qu, J., Snider, B.J., Wu, S., 2014. Proteasome inhibitors

Nakamura, Y., 1998. Isolation and characterization of a novel human lung- activate autophagy involving inhibition of PI3K-Akt-mTOR pathway as an anti-

speciﬁc gene homologous to lysosomal membrane glycoproteins 1 and 2: oxidation defense in human RPE cells. PLOS ONE 9, e103364.

signiﬁcantly increased expression in cancers of various tissues. Cancer Res. 58, Thal, D.R., Rub, U., Orantes, M., Braak, H., 2002. Phases of A beta-deposition in

3499–3503. the human brain and its relevance for the development of AD. Neurology 58,

Parganlija, D., Klinkenberg, M., Dominguez-Bautista, J., Hetzel, M., Gispert, S., Chimi, 1791–1800.

M.A., Drose, S., Mai, S., Brandt, U., Auburger, G., Jendrach, M., 2014. Loss of PINK1 Vogl, D.T., Stadtmauer, E.A., Tan, K.S., Heitjan, D.F., Davis, L.E., Pontiggia, L., Rangwala,

impairs stress-induced autophagy and cell survival. PLOS ONE 9, e95288. R., Piao, S., Chang, Y.C., Scott, E.C., Paul, T.M., Nichols, C.W., Porter, D.L., Kaplan,

Park, C., Cuervo, A.M., 2013. Selective autophagy: talking with the UPS. Cell Biochem. J., Mallon, G., Bradner, J.E., Amaravadi, R.K., 2014. Combined autophagy and pro-

Biophys. 67, 3–13. teasome inhibition: a phase 1 trial of hydroxychloroquine and bortezomib in

Pennati, M., Lopergolo, A., Profumo, V., De Cesare, M., Sbarra, S., Valdagni, R., Zaf- patients with relapsed/refractory myeloma. Autophagy 10, 1380–1390.

faroni, N., Gandellini, P., Folini, M., 2013. miR-205 impairs the autophagic ﬂux Wilke, S., Krausze, J., Bussow, K., 2012. Crystal structure of the conserved domain of

and enhances cisplatin cytotoxicity in castration-resistant prostate cancer cells. the DC lysosomal associated membrane protein: implications for the lysosomal

Biochem. Pharmacol. glycocalyx. BMC Biol. 10, 62.

Pihlstrom, L., Axelsson, G., Bjornara, K.A., Dizdar, N., Fardell, C., Forsgren, L., Holm- Zang, Y., Thomas, S.M., Chan, E.T., Kirk, C.J., Freilino, M.L., DeLancey, H.M., Gran-

berg, B., Larsen, J.P., Linder, J., Nissbrandt, H., Tysnes, O.B., Ohman, E., Dietrichs, dis, J.R., Li, C., Johnson, D.E., 2012. The next generation proteasome inhibitors

E., Toft, M., 2013. Supportive evidence for 11 loci from genome-wide association carﬁlzomib and oprozomib activate prosurvival autophagy via induction of the

studies in Parkinson’s disease. Neurobiol. Aging 34 (170), e1707–e1713. unfolded protein response and ATF4. Autophagy 8, 1873–1874.